You are on page 1of 42

Due Diligence Guidelines

For The Review of Microcredit Loan Portfolios

A Tiered Approach
2009

Robert Peck Christen and Mark Flaming

These guidelines were first developed by Robert Christen, with assistance from Patricia Mwangi,
both of CGAP. Mark Flaming drafted the current version after testing the methodology. Richard
Rosenberg of CGAP wrote the annex on yield gap analysis. Lynn Curran and Deborah Drake of
ACCIÓN, Todd Farrington of MicroRate, and Mark Schreiner of Washington University
contributed to earlier versions of a related document.

The authors and CGAP would be very grateful for feedback from people or
organizations that have used this tool in testing the quality of a microloan portfolio.
Please send any comments on your experience with it to rrosenberg@worldbank.org.
Your observations will help us improve future editions of the tool.

ABSTRACT
Microfinance Institutions (MFIs) operate with risks that investors need to be concerned about.
Unfortunately, external audits, ratings, evaluations, and even supervision too often fail to identify
the primary risk – inaccurate representation of portfolio quality. These due diligence guidelines
are designed to help an analyst evaluate, with statistical precision, the accuracy of accounting
and performance reports about the portfolio, as well as an MFI’s compliance with its own
portfolio management policies. The document guides the analyst through three increasingly
robust levels of appraisal. This allows the analyst to adapt the exercise to a wide range or uses
and requirements about levels of confidence. In the third level, the use of statistical sampling to
validate conclusions from conventional appraisal techniques in the prior levels generates
uniquely robust results that should satisfy the most demanding commercial investor or regulator.

Tier I is a five-person-day review of the MFI’s basic policies, procedures and systems for
managing and reporting on the performance of its loan portfolio. This review should be required
as a part of any credible external assessment, appraisal, or ratings exercise.

Tier II is a 10 – 14 person-day assessment of whether operational practices are consistent with
the MFI’s policies and procedures and with standards of best practice in microcredit portfolio
management. This would normally be a minimum requirement for donors, investors, and ratings
agencies that evaluate and assess MFIs for potential investment.

Tier III is a 3 - 4 person-week exercise that includes an audit to confirm the portfolio quality
through a sampling of loan files, accounting files, and the loan tracking software. Tier III due
diligence is recommended for investors or regulators who need a quantitative and statistically
certain measure of portfolio performance.

These due diligence guidelines provide a comprehensive checklist of what to analyze, some
instruction about appraisal technique, and very little guidance about how to interpret the results.
The Tier I review can be carried out by someone with general knowledge of microfinance but
Tiers II and III rely heavily on the judgment of an analyst with extensive experience in the
evaluation of MFIs and their loan portfolios.

For another microcredit portfolio testing tool, produced by MicroSave, see
http://microsave.org/relateddownloads.asp?id=14&cat_id=313&title=Loan+Portfolio+Audit
Due Diligence Guidelines for the Review of Microcredit Loan Portfolios

INTRODUCTION
Microfinance Institutions (MFIs) operate with risks that investors need to be concerned about.
Unfortunately, external audits, ratings, evaluations, and even supervision too often fail to identify
the primary risk – faulty representation of portfolio quality. These due diligence guidelines are
developed to help investors, donors, and regulators verify the real level of risk in an MFI’s loan
portfolio.

Discrepancies between real and reported portfolio quality are common in MFIs. In fact, many
well-known MFIs have experienced at least one significant portfolio crisis, sustaining
delinquency and default rates well above what they reported to the public. Finansol, for example,
was a Colombian MFI that was widely considered to be one of the top two or three institutions in
Latin America during the early part of the 1990s. In November 1997, the Colombian bank
superintendent required Finansol to inject several million dollars of capital to shore up reserves
for bad debt. The intervention came after inspectors reclassified a significant number of
Finansol’s loans as “doubtful.” The additional reserve requirement put Finansol in technical
insolvency, since the associated expenses decreased capital below acceptable adequacy levels.

Finansol’s case was highly public, but certainly not unique. Because it was a supervised financial
institution its problems were a matter of public record. But there have been many cases that did
not receive publicity. Most microfinance network organizations have discovered--more than
once--that one of their most important affiliates had a far more serious problem with loan
delinquency than previously stated. These incidents have cost many millions of dollars, and in
some cases have forced donor agencies to recapitalize or shut down operations they have
supported for years.

The occasional failure of MFI managers, donors, or investors to detect portfolio failures is
simply part of the growing pains of a generally vibrant microfinance industry. Fortunately, most
MFIs have been able to weather these episodes because they were only moderately leveraged,
their portfolio problems did not create an acute liquidity crisis, or they were able to find another
donor to recapitalize the institution. However, MFIs are integrating into regulated financial
systems, taking deposits, and leveraging themselves with commercial sources of funding. The
Finansol case demonstrated that poor oversight can lead to severe consequences in these
conditions.

The purpose of these due diligence guidelines is to provide banking superintendents, donor
agencies, and potential investors with a tool for a more accurate and deeper understanding of the
underlying quality of an MFI’s loan portfolio than current methods offer. This tool is designed to
create the levels of confidence required for MFIs to leverage their capital with commercial
funding and deposits.

These due diligence guidelines provide a comprehensive checklist of what to analyze, some
instruction about appraisal technique, and very little guidance about how to interpret the results.
This is not a training manual for aspiring analysts; it is a checklist for an analyst with
extensive experience in the evaluation of MFIs. Earlier versions of this document attempted to
annotate the checklist with guidance about how to interpret the results of the appraisal, but these
proved to be unsatisfactory and cumbersome in practice. This version relies ultimately on the

Page 1
Due Diligence Guidelines for the Review of Microcredit Loan Portfolios

judgment of the analyst to use the checklist to sort out what matters most and to identify the root
causes of risk in the loan portfolio.

THE DUE DILIGENCE PROCESS
Due diligence is the process of examining the performance of a company. It is typically
performed by or on behalf of investors who are primarily interested in verifying that performance
is accurately reflected in the financial statements and company reports, and that management and
operating systems are robust enough to sustain good performance into the future.

In the commercial financial industry, investors rely heavily on external appraisals of a financial
institution to analyze performance trends. Audits, ratings reports, stock analyst reports,
supervisory reports, and even media articles provide an array of appraisals of a financial
institution’s performance. This varied and ongoing measurement of performance provides
incentive to the financial institutions, creates consensus around best practice metrics, and
provides an evaluator with many sources of analysis. Thus the due diligence analyst typically
arrives at a financial institution with considerable information. The due diligence process itself
typically entails a review of the basic management and operational systems and a careful analysis
of the financial institution’s business model. The important point to make about due diligence
methods in a commercial and regulated financial industry is that the analysts typically rely to
a great extent on the information provided by management and by the accounting and
performance reporting systems.1 They do this based on the assumption that an institution that
operates successfully in such an industry has robust systems and competent management, and
that gross deficiencies tend to get weeded out by competition, audits, supervision, or external
appraisals.

This assumption is not always valid in the microfinance industry, so conventional approaches to
audits or rating appraisals can fail to capture the peculiar weaknesses of MFIs. The microfinance
industry has placed great emphasis on adapting appraisal and evaluation tools, audit and financial
disclosure guidelines, and performance standards. There are assessment tools, such as ACCIÓN
International’s CAMEL, WOCCU’s PEARLS, and CGAP’s Technical Tool for appraising MFIs;
assessments by rating agencies (MicroRate, M-CRIL, Microfinanza, CRISIL, and PlaNet
Finance); credit ratings by credit rating agencies (Standard and Poor’s, Fitch, Apoyo, Class,
Pacific Credit Rating, and Equilibrium); guides for analysts, including the Technical Guide of
the Inter-American Development Bank, CGAP’s external audit manual, and minimum disclosure
guidelines for MFI financial statements. There are also benchmarking projects, such as the
MicroBanking Bulletin and MIX Market. In addition, there are a whole host of local or national
efforts to develop many of the same tools and capacities to support network reporting
requirements, apex investment decisions, and inspection visits by bank superintendents.

These appraisal tools are effective at assessing an MFI’s overall financial performance. But they
typically do not test the veracity of the underlying information. Appraisal work plans typically
focus on loan portfolio administration without delving much into a detailed reconciliation of loan
file documentation, accounting records, and loan tracking system reports. The appraisals
typically assume that the information that is reported is correct, and make that assumption based
1
In researching due diligence practices, the authors interviewed many representatives from ratings agencies,
investment banks and commercial funders who confirmed that information risk is not typically addressed as a
primary concern in traditional due diligence practice.
Page 2
Due Diligence Guidelines for the Review of Microcredit Loan Portfolios

on the existence of an acceptable audit report. Conventional audits usually include some tests of
these systems, but they are seldom intensive enough to justify much confidence in the accuracy
of system reports, particularly loan tracking reports.

The following example demonstrates this point. In a typical audit, letters are sent out to a random
sample of loan clients asking them to verify the loan they received, their outstanding balance,
their latest payment, and any other pertinent information. Clients are asked to send the letter back
to the auditor if they disagree with any of the information provided in the letter. The auditor
assumes that if no letters come back, there are no disagreements. The flaws in this assumption
are of course obvious to anyone with experience with the typical clientele of an MFI.

Due diligence in an MFI must go farther than conventional methods to verify performance
precisely in the areas that typically challenge MFIs. These due diligence 2 guidelines are designed
for that purpose.

A THREE-TIERED APPROACH
This document guides the analyst through a three-tiered approach for assessing how well an
MFI’s loan portfolio is performing and how well it is managed. The tiers build on each other,
each employing more sophisticated analytical procedures and more evaluation staff resources
than the previous tier. A due diligence process that incorporates all three tiers offers the greatest
precision and highest degree of certainty about the actual level of risk in the loan portfolio.

Tier I: This is a five-day review of the MFI’s basic policies, procedures and systems for
managing and reporting on its loan portfolio. It is both a high-level analytical review of portfolio
trends and a stock-taking exercise. In the stock-taking exercise the analyst collects the MFI’s full
inventory of credit policies, procedural manuals, key accounting policies, financial and
performance reports. The Tier I assessment is based largely on the completeness and coherence
of this inventory and on interviews with management about portfolio trends and the policies,
procedures and systems for managing the loan portfolio. The collection of documents also
provides the basic information for Tier II analysis.

Tier II: This is a 10 – 14 day assessment of whether MFI operational practices are consistent
with policies and procedures and with standards of best practice in microcredit portfolio
management. The analyst verifies that accounting policies and reporting formats are consistent
and based on sound practices; that credit department staff conduct operations according to policy
and procedural guidelines; and that the MFI does not regularly engage in practices that generate
unmeasured credit risk, such as reclassifying loans, refinancing, and granting parallel loans. The
Tier II assessment is based primarily on how closely the MFI’s practices match internal policies
and how reports match industry standards of best practice.

Tier III: This is a 3 to 4 week exercise that includes detailed testing of transactions to confirm
the portfolio quality through a sampling of loan files, accounting files, and the loan tracking
MIS. The Tier III audit employs statistical sampling methods that support conclusions, within
acceptable levels of confidence and error, about real arrears levels and, more broadly, the risks

2
These due diligence guidelines deal specifically with portfolio quality and management. A complete due diligence
process in an MFI would involve the many other areas that are typically included in a full due diligence exercise.
Page 3
Due Diligence Guidelines for the Review of Microcredit Loan Portfolios

related to accounting practices and credit policy. The Tier III conclusions are supported by the
results of the audit, namely, whether portfolio performance is accurately reflected in accounting
and performance monitoring reports.

The Tier I review should be required as a part of any credible external assessment, appraisal, or
ratings exercise. It will give the analyst an overview of recent company trends that may affect the
loan portfolio and a basic but comprehensive look at the MFI’s policies, procedures, and reports.
The evaluation incorporates enough discussion, cross references of key performance indicators,
and accounting and MIS review to provide a sense of whether or not the MFI at least has the
basic tools to manage a portfolio. The analyst will make qualitative and subjective judgments in
a Tier I review, so the value of the conclusions will depend greatly on the analyst’s level of
expertise. It is often sufficient to detect glaring deficiencies. However, investors that require
more robust evidence of actual performance will need to proceed at least to Tier II.

A Tier II review is a minimum requirement for donors, for investors, and for ratings agencies
that evaluate MFIs for potential investment. The Tier II analysis confirms that the MFI’s
policies, procedures, systems and reports comply with minimum industry standards and that they
are in fact employed in MFI operations down to the branch office level. Tier II conclusions about
MFI performance and overall management and operational capacity are supported as well by an
assessment of the accounting and portfolio tracking systems. The Tier II appraisal also draws on
the subjective judgment of the analyst. However, the extensive interviews and systematic cross-
checking of policies and actual practice provide the analyst with multiple channels of
investigation into all parts of the organization.

Tier III due diligence is recommended for investors or regulators that need a quantitative and
statistically certain measure of portfolio performance and management. This would be banking
superintendents, investors considering a significant equity investment, or an investment bank
preparing the MFI for accessing capital markets. Tier III provides statistically significant tests of
loan quality and the integrity of reporting and internal control systems

The incremental approach to due diligence enables the potential investor or donor to confine the
analysis to the level of certainty they need, thereby avoiding unnecessary time and expense. The
analysis can even be halted when the desired level certainty is obtained. If, for example, a Tier I
examination reveals incoherent credit policies and untransparent accounting practices, there may
well be no point in a strict statistical analysis of the full portfolio. For some investors, there will
be no need to extend the analysis beyond the point that significant portfolio problems are
detected. However, should it be necessary to quantify more precisely the extent of the arrears,
refinancing, or fraud, the analyst can employ the next level of analysis.

CONTRACTING GUIDELINES
The due diligence sponsor is advised to follow three basic guidelines for contracting an analyst
to use this methodology:
• select an analyst, or analyst team, with the requisite skill set;
• work with the analyst to define the scope of work and workplan; and,
• clearly define expectations about what the sponsor needs to know to make its decision, and
the desired levels of precision when supporting conclusions based on statistical sampling.

Page 4
Due Diligence Guidelines for the Review of Microcredit Loan Portfolios

Contracting for a Tier I exercise should be fairly straightforward. The scope of work is already
explicit in these guidelines and will not vary significantly.

The Tier II guidelines also provide a comprehensive scope of work. However, the Tier II
exercise is more dynamic in that the analyst must be able to direct efforts towards problem areas
as they are discovered. The sponsor and analyst should develop a workplan that combines the
Tier I and Tier II instructions, and that anticipates mid-course adaptation at key points. The first
checkpoint will be just prior to the field visit when the analyst knows the condition of the MFIs
basic operating documents. The second will be after the analyst has concluded the interviews
with headquarters management staff. At this point, both the sponsor and the analyst should
expect to review and adapt the remaining workplan based on initial findings.

The Tier III exercise requires careful selection of the analyst team and definition of the scope of
work. These elements will vary depending on the objectives of the sponsor and the
characteristics of the MFI. The lead analyst or analysts must have the capacity to organize the
Tier III procedures to test the conclusions of the Tier II assessment. This requires extensive
experience with MFI management and evaluation. The Tier III team must also have the requisite
skills in financial and statistical analysis, spreadsheet use, and auditing procedures. The most
feasible way to construct such a team is to hire a lead analyst, who in turn hires supporting
analyst and/or a local auditing firm that is familiar with the microfinance industry and with
statistical sampling methods. As in the Tier II exercise, the scope of work should anticipate
review and adaptation of the workplan at the key checkpoints. In Tier III, this is especially
important as the analyst makes decisions about how to sample the portfolio in the audit. The
analyst, audit sponsor, and the MFI are encouraged to design the sample method with clear
expectations about the levels of certainty in the results.

Due diligence sponsors should make special effort to clarify all expectations when working with
audit firms. The three-tiered approach is more extensive and more detailed than the scope of
conventional audits. The sponsor will want to sign a contract with the auditor that specifies both
the procedures and the auditor’s responsibility for formulating conclusions on the results. Many
auditors are disinclined to use any but the standard procedures employed by their firm. Even
firms that agree to conduct the procedures may not be willing to issue an opinion on their results.
In these circumstances, sponsors will do well to hire a lead analyst who is responsible for
formulating final conclusions and for subcontracting and managing the audit firm.

Sponsors that require a precise valuation of the portfolio may need to contact an audit firm and
discuss the level of sampling required to achieve the desired results. The design and cost of such
an audit will likely surpass the typical scope of a Tier III audit. Statistical conventions for
population stratification and sample sizes will require fairly large samples by branch office,
lending product, client type, and even credit officer. Large audit firms have their own way of
designing samples and will probably insist on using their internal systems as a condition for
issuing an official opinion.

Page 5
Due Diligence Guidelines for the Review of Microcredit Loan Portfolios

1. TIER I DUE DILIGENCE

The initial level of portfolio due diligence encompasses what should be the first step in any
assessment of an MFI’s performance, credit operations and accounting practices. The primary
objective of the Tier I review is to determine if the MFI has adequate policies, operating systems,
monitoring, and management capacity to sustain portfolio quality. The analyst will collect
documents and reports, conduct a basic review, and interview top management about the basic
business model of the MFI and recent trends. This assessment should not require more than two
days in the head office, plus preparation time of approximately three days. 3

Potential investors and donors should be aware that the Tier I assessment does not attempt to
determine whether the MFI complies with its own policies, or whether its reporting systems are
accurate. However, a Tier I assessment will provide a basic understanding of portfolio
management that is considerably better than the results achieved through a mere desk review of
reports and documents. The on-site interviews with senior management, combined with a prior
desk review, are often sufficient to detect material deficiencies, or to confirm the consistency and
comprehensiveness of the MFI’s basic organizational infrastructure.

The Tier I process will typically consist of four steps:
• gather information prior to field visit;
• desk review;
• field visit;
• prepare final analysis.

1.1. Information Requirements
The process of gathering information is the lengthiest step in Tier I due diligence and it will
provide the analyst with important clues about the capacity of the MFI to manage its portfolio
operations. The analyst should gather as much information as possible from sources both inside
and outside the MFI. The following is a summary of the required documents (See Section 1
tables in Annex I for additional discussion of document content).

Financial reports.
• audited financial statements (balance sheet, profit and loss, and cash flow) for the past
two complete fiscal years. The audited financials serve as a benchmark for the other
sources of information in the desk review, as they represent an outside validation of the
information provided by MFI management. The audited financial statements report
should consist of an auditor’s opinion, financial statements, and notes to the financial
statements. The analyst should ask for the complete report, including the management
letter if the auditor provided one, and the MFI management’s response to the auditor’s
report; 4
• current year-to-date financial statements if more than one quarter has elapsed;

3
The amount of time required to collect the necessary documents from the MFI prior to the trip can vary greatly. If
documents are not collected prior to the field visit, more in-country days will be required.
4
For further detail on the audit of MFIs, see External Audits of Microfinance Institutions: A Handbook, CGAP
Technical Tool No. 3 (Washington, D.C.: CGAP, December 1998), especially pages 22-23.
Page 6
Due Diligence Guidelines for the Review of Microcredit Loan Portfolios

• supporting internal financial reports for the same periods that provide sufficient detail of
the loan portfolio (classification by risk/reserve categories), reserves, income and
expenses;
• a current liquidity and asset/liability management report.

Accounting policies. This will consist of an accounting policy manual and memos related to the
following transactions:
• accounting of loan balances by arrears classification;
• rescheduled loans;
• loan loss provisions 5, loan loss provision expenses, loan write-offs, and recoveries (it is
particularly important to get a clear statement of the MFIs policies in these areas);
• fee and interest income on loans, including the accrual of interest;
• allocation of client payments to outstanding balances;
• liquidation of collateral;
• any non-cash methods of loan repayment (e.g., in-kind, refinancing old loans with new
ones).

Credit policies and procedures. The MFI should have documents that provide clear and
comprehensive guidelines for portfolio and delinquency management. These documents should
include any credit-related policies and procedures manuals or memos to staff.

Portfolio reports. The loan portfolio reports are typically produced by the MFI’s management
information system (MIS) 6 and provide detailed information about lending activities. The MIS
should be able to produce the following reports for the analyst:
• portfolio classification of loans and amounts by number of days delinquent;
• a comparison of actual loan repayments to scheduled repayments (during the past year or
quarter);
• breakdown of portfolio by loan product (lending methodology), including number of
active loans, total portfolio, and portfolio quality for each category;
• portfolio reports by branch and/or loan officer (including portfolio size, details on
products, portfolio quality);
• detailed reports on restructured and refinanced loans; and
• summary portfolio performance reports for branch managers, senior management, and the
board of directors.

Assessments, appraisals, and external evaluations. The MFI should send the analyst the report
generated by any assessment, appraisal, or external evaluation that has been done during the
prior two years. This information is generally confidential and should be treated as such by the
analyst. Increasing numbers of these reports are being prepared by ratings agencies for affiliate
networks, donors, and investors.

5
Sometimes refered to as impairment allowance, impairment expense or impairment loss.
6
For further information on the development of quality management information systems for MFIs, see
Management Information Systems for Microfinance Institutions, CGAP Technical Tool No. 1 (Washington, D.C.:
CGAP, February 1998).
Page 7
Due Diligence Guidelines for the Review of Microcredit Loan Portfolios

1.2. Desk Review
The analyst should prepare for the site visit with a desk review organized around three
objectives. One objective is for the analyst to gain an understanding of the business model and
operations of the MFI, including the most recent growth, provisioning and write-off trends. This
will guide the analyst’s interviews with senior management. The analyst should also discuss any
major changes in the market or in MFI operations and determine how these might affect portfolio
management. A second objective is to determine if the MFI has adequately documented policies,
procedural manuals, and financial and portfolio performance reporting. The ability of the MFI to
provide these documents is the first indication of the management capacity and professionalism
of the institution. Anything short of immediate delivery of a complete set of documents and
reports should guide the analyst towards areas that require special attention during the field visit.
And the third objective of the desk review is to identify policies or performance that deviate
substantially from best practices of similar MFIs.

The Section 1 tables in Annex I serve as an appraisal guide for the analyst. The main components
of the analysis are summarized below:

The MFI Operating Environment. Most of this information will be gathered from interviews
with senior management during the field visit. However, the analyst should gather information
from www.themix.org to compare the MFI to competitors in the market (or similar MFIs
elsewhere) and to general market trends.

Financial Reporting and Performance. The analyst should assess the overall financial
performance of the MFI as well as the quality of the underlying information.
• Accounting policies and financial reports
It is very important to confirm that accounting policies and report formats use accepted
standards for the key variables that reflect portfolio performance (e.g., interest accrual,
arrears classification, loan loss provisioning, write-offs, etc.).
• Reporting consistency
It is important to verify that the amounts are consistent between reports, especially the
portfolio balances reported in the loan tracking system and on the balance sheet.
Discrepancies are common because data often comes from different sources or data flows.
The analyst should discuss discrepancies with MFI management to determine whether
management actively reconciles discrepancies and whether the discrepancies represent a
material deficiency in the reporting system.
• Financial Performance: Assets
The financial analysis should focus on the evolution of key indicators over time, looking
for abrupt changes or performance that deviates from normal standards. Key indicators
include:
o growth rates of loan disbursements and portfolio balance;
o portfolio in arrears;
o loan loss provisions; and,
o write-offs.
• Financial Performance: Income and Expenses

Page 8
Due Diligence Guidelines for the Review of Microcredit Loan Portfolios

o yield on portfolio, including a yield gap analysis that compares actual with
expected income 7;
o cost of funds;
o provisioning and write-off expenses as a percentage of the average loan portfolio;
o operating expenses.

The analyst will be guided to a great extent by his or her own judgment about the indicators that
best reveal portfolio quality in each particular case. There may be good reason to make use of
additional indicators of financial performance. 8

The analyst will also want to consider the MFI’s performance against peer institutions. MFIs
differ widely in the clients they serve, their methods of managing portfolio risk and delinquency,
their institutional structure, and the legal and regulatory environments in which they are operating.
In addition, contextual factors such as geographic location, local market conditions, and age of the
institution influence performance indicators. In recent years, the MicroBanking Bulletin (MBB)
has begun to address the problems of peer group analysis. The MBB tracks the financial
performance of over 700 microfinance institutions from around the world (www.mixmbb.org).
Although the individual program results are kept confidential, the MBB displays results for peer
groups of similar institutions, which the analyst can use as a guide to determine where the
performance of given institutions fall. In addition, affiliate networks, such as ACCIÓN
International, FINCA, CASHPOR, MFN, and SEEP, may have internal databases on which to
draw peer comparisons. Finally, the analyst can look at the performance of other individual MFIs
in the country at the MIX Market (www.mixmarket.org).

Credit Policies and Procedures
A clearly documented credit policy is a necessary condition of prudent lending. The MFI’s
lending policies should establish the necessary authority, rules, and framework to operate and
administer its loan portfolio and manage risk. The policies should be appropriate to the size of
the institution and the nature of its activities, and should be consistent with prudent banking
practices and relevant regulatory requirements. The policies should be broad and not overly
restrictive in setting guidelines to maintain sound credit appraisal and approval standards and
control, manage risk, and deal with problem loans. 9

The desk review of credit policies and procedures focuses on four key questions:
• Are the eligibility criteria and appraisal and approval procedures adequate and clearly
documented to ensure consistent loan analysis?
• Do the application and approval forms reflect the loan analysis and approval procedures?

7
This quick and very effective check on whether there are hidden arrears in the portfolio is discussed in Annex II.
8
There are a number of publications that explain the indicators commonly used in the microfinance industry: See
Tillman Bruett, ed., Measuring Performance of Micro-finance Institutions: A Framework for Reporting, Analysis,
and Monitoring, SEEP and ACT (Washington, DC: USAID/Accelerated Microfinance Advancement Project,
forthcoming); CGAP, Selected Definitions of Financial Terms, Ratios and Adjustments for Microfinance,
Microfinance Consensus Guidelines (Washington, DC: CGAP, 2003); Joanna Ledgerwood, “Sustainable Banking
with the Poor” in Microfinance Handbook: An Institutional and Financial Perspective (Washington, DC: IBRD/The
World Bank, 1998); CGAP, Measuring Microcredit Delinquency, Occasional Paper #3. (Washington, DC: CGAP,
1999.
9
Board of Governors of the Federal Reserve System, Commercial Bank Examination Manual. Division of
Supervision and Regulation (Washington D.C., March 1994).
Page 9
Due Diligence Guidelines for the Review of Microcredit Loan Portfolios

• Is the approval process structured with adequate levels of authority and control? And,
• Does the MFI have loan policies commonly associated with poor performance?

Portfolio Reporting
The desk review of portfolio reports is primarily a stock-taking exercise. The analyst reviews the
key reports to determine whether the MIS generates a complete set of reports that cover all
necessary areas of credit operations, and whether the balances match the financial reports.

Management Control of Portfolio Performance
Even in a Tier I exercise, the analyst should be able to find evidence of management oversight
and control of portfolio performance. The level of detail and frequency of portfolio reports, and
the manner in which they are actually used, will provide important clues. The analyst should pay
close attention to level of development of internal control and audit systems.

1.3. The Field Visit
The field visit provides the analyst with an opportunity to gather missing information and ask
clarifying questions, to check for consistency between MFI policy and management’s
understanding of those policies, and ultimately to form an opinion about senior management
capacity.

The desk review should prepare the analyst with a list of missing documents and reports as well
as questions around key issues. It is not uncommon for the analyst to devote a significant share
of the field visit to simply completing the document inventory and asking clarifying questions
about the policy documents and reports. The analyst may need to spend time with junior staff in
the organization to understand exactly how key ratios relating to loan portfolio performance are
generated and the basis for their calculation. 10 Managers rarely understand the exact manner of
their calculation, even if they think they have given clear orders for how they should be
calculated. It is worth stating again that the analyst may learn the most about the MFI from
mangers’ ability to produce a complete set of these documents and reports and explain how they
are used.

The on-site visit should begin with a meeting with the senior management team of the institution.
The meeting should include discussions of the economy, regulatory environment, the state of the
microfinance market in the country, and the institution’s clients. Through these discussions, the
analyst should gain an understanding of the context in which the MFI operates and disburses
credit.

The analyst should then hold subsequent meetings with members of the management team to
discuss the MFI’s loan portfolio performance and management. These face-to-face meetings
should cover all of the policies and practices of the institution that are listed in the Section 1
tables in Annex I. This process is iterative in that the analyst will ask similar questions of the
general manager, chief financial officer, the credit manager, and credit officers of the head

10
For further information on the various methods of measuring delinquency and the pros and cons of each indicator,
see Rosenberg, Measuring Microcredit Delinquency: Ratios Can Be Harmful to Your Health, CGAP Occasional
Paper, no. 3 (Washington, D.C.: CGAP, 1999).
Page 10
Due Diligence Guidelines for the Review of Microcredit Loan Portfolios

office. The answers to these questions should be consistent among all members of the team, and
should correspond to the stated policies of the institution.

Analysis of credit policy should be grounded in discussions with both management and credit
department staff, and with reference to the credit manuals. The analyst must look at the
principles underlying an MFI’s credit methodology in order to answer this essential question:
Are there clearly defined and consistently applied credit policies? There is frequently a
difference between what is printed and what is practiced. For this reason, what management and
staff say and do is always more important than what is printed in the manual.

1.4. Final Analysis
The analyst needs to form an overall impression of the portfolio risk of the MFI. Much of this
impression will depend on the analyst’s confidence about the trustworthiness of the information
and the senior management team. The analyst should consider the transparency of data, the ease
with which it is provided, the familiarity of top managers with loan-portfolio quality indicators
and reports, the consistency of information across several MIS reports, the commitment of staff
to ironing out inconsistencies, the general environment of orderliness and internal control, the
presence of actual ex-post control mechanisms, and clarity in all aspects of the MFI’s operation.
When an analyst concludes the Tier I exercise with an uncomfortable feeling, chances are high
that the results of Tiers II and III exercises would be substantially worse. Rarely do microcredit
portfolios look better upon closer examination than they do at the outset.

2. TIER 2 DUE DILIGENCE
Tier II due diligence builds on the Tier I exercise. 11 After completing all of the tasks for Tier I,
that analyst completes five additional areas of analysis:
• the integrity and completeness of the MIS;
• a verification of loan classification, provisioning and write-off procedures;
• the actual performance of credit operations, especially in the branch offices;
• the degree to which the work of individual loan officers reflects MFI policy; and
• the integrity and completeness of individual loan files.

The essential question of Tier II is whether or not the policies and procedures presented by senior
management and in MFI documents is reflected in practice at the operational level, and, whether
those operational practices are sound. In Tier II, the analyst engages the staff in the credit,
accounting and information technology departments and, most importantly, the branch offices.
The process will be iterative, sometimes requiring the analyst to consult with several departments
to confirm how transactions are recorded, how reports are generated, and how all loan
management procedures are actually carried out.

A Tier II exercise may be more effective if conducted by a team of two analysts, for example one
with financial analysis skills and the other with credit operations experience, to bring a wider

11
For the sake of simplicity, the remainder of this section will refer only to the additional work required in a Tier II
exercise. It is also important to clarify that a Tier I analysis that is conducted in a Tier II due diligence exercise will
usually be more robust than if carried out alone. Although the analyst asks the same questions in both cases, in a
Tier II exercise the analyst has more opportunities to collect, verify, and analyze information.
Page 11
Due Diligence Guidelines for the Review of Microcredit Loan Portfolios

range of experience to the task. The analysts are likely to benefit from comparing perspectives
and discussing findings.

2.1. MIS Review
In Tier Two due diligence, the analyst will focus on verifying the accuracy of reports and the
capabilities of the accounting and the portfolio tracking systems. The Tier I review will likely
guide the analyst towards the areas of these systems that merit further study. A lack of clarity
around accounting practices, or discrepancies between reports, provide important clues for
further investigation.

The most effective method for testing accounting practices and reporting accuracy is to audit the
transactions on a sample of loans, following each transaction through the portfolio tracking and
accounting system. The analyst will need to first develop a spreadsheet template that replicates
all of the transactions associated with a single loan: disbursements, fee assessments, interest
calculation, distribution of principal, interest and fee payments, arrears calculation, etc. The
formulas must precisely reflect the policies and procedures for these transactions. The analyst
then selects a sample of loans from the accounting system, collects the corresponding contracts,
and prints a report from the portfolio tracking system that has a full transaction history for each
loan. The analyst enters the disbursement and payment transaction information from the loan
report manually into the spreadsheet template so that the spreadsheet formulas calculate the
running balances, principal, interest and fee allocation, and arrears status. The analyst then
compares the result of the spreadsheet template to the reports generated by the portfolio tracking
system.

The analyst then confirms that the transactions are passed accurately into the accounting system.
The analyst should also inspect the physical receipts associated with the disbursements and client
loan payments. This completes a full audit of the transactions that are reflected in the balances
that appear in the financial statements and portfolio reports.

In this exercise, the analyst will confirm whether the portfolio tracking system (i) reflects the
exact terms and conditions of the individual loan contracts (ii) accurately calculates and reports
principal, interest and fee distributions, (iii) accurately calculates the arrears classification of the
loans, (iv) produces accurate reports, and (v) passes the information accurately to the accounting
system.

The analyst should conduct the transaction audit on a sample of loans that includes examples of
all of the MFI’s lending products. This sampling technique makes no attempt at statistical
significance. It simply helps the analyst understand the details of the loan products, focuses
discussion with staff, and detects systematic deviations from stated policies.

The analyst should also comment on whether the accounting and loan tracking system are
capable of expanding with the MFI. The challenges and potential of these systems vary
significantly. Custom designed software, off-the-shelf software, or systems purchased from other
MFIs require different skill sets to operate and maintain. The analyst should assess whether the
MFI has the resources, or access to the resources, to maintain the system it is using.

Page 12
Due Diligence Guidelines for the Review of Microcredit Loan Portfolios

2.2. Loan Classification, Provisioning and Write-offs
The Tier II analysis should evaluate the MFI’s method for classifying its portfolio, estimating
and provisioning for loan losses, and writing off and recovering bad loans. The evaluation should
compare real losses to provisioning levels. For most MFIs, the risk classification system should
be based on repayment performance as well as historical patterns associated with factors such as
specific lending products, guarantee coverage, economic sectors, geographic location or branch
office. MFIs with finely tuned classification criteria use classifications not only for provisioning
but also to set differential risk-based interest rates. 12

2.3. Branch Office Management
This part of the Tier II exercise focuses on how well the MFI manages credit operations through
the branch offices. Most of the information for this analysis is gathered through interviews with
credit department management and branch office staff. The analyst should structure the analysis
around five basic areas.

Distribution of authority and responsibility. The analyst should map how the key credit
activities and decisions are organized between headquarters and the branch offices. Ultimately,
the analyst will be looking for an appropriate balance of operational efficiency and control in the
MFI’s management of lending targets, loan approvals, cash management, accounting and
reporting.

Performance management. The analyst should pay special attention to how branch office
performance is managed. The mapping exercise should reveal how responsibility is distributed
for setting performance targets, gathering performance indicators, performance reporting,
performance review, and management response to performance achievements. The analyst
should expect to find robust systems for measuring performance and decisive management
response to both good and poor results.

Core operational policies. The operations most directly related to portfolio quality are loan
analysis and delinquency management. The interviews should reveal staff understanding and
competency in these areas.

Consistency between policies and practice. The analyst should pay special attention to the
consistency between how things are done at the branch and how headquarters management
thinks things are done, in all areas of credit operation. Discrepancies reflect lack of management
control.

Branch office morale. Staff morale is an important indicator of current and future portfolio
performance. Low morale is often a product of staff frustration over ineffective policies,
inadequate management, poor compensation, or chronic poor portfolio performance. Sustained
portfolio performance is unlikely with demoralized staff.

2.4. Loan Officer Performance
Loan officers are one of the most important sources of information about lending operations.
They carry out most of the procedures and they are closest to the clients. The analysts should

12
The ACCIÓN CAMEL, Technical Note, Microenterprise Best Practices, p. 42.
Page 13
Due Diligence Guidelines for the Review of Microcredit Loan Portfolios

expect to spend at least three full days with the loan officers, away from senior management, in
order to capture the reality of field operations. The analyst should use informal conversations
with loan officers to cover the topics of portfolio quality, how closely lending policies are
followed, how flexible they are, routine contact with current clients, loan
rescheduling/refinancing, and delinquent loan procedures. The analyst should be careful how
they question loan officers in order to avoid intimidating field staff and impeding the
discussions.

The analysts should also accompany loan officers on client visits to witness first-hand how
procedures are followed in the field. On these visits, the analysts should be able to get a sense of
the client’s history with the MFI, level of satisfaction, relationship with loan officer, etc.
Discussions with clients should be brief and unstructured with a few open-ended questions.

The analyst should also review the MFI’s training program for credit officers and any incentive
systems to encourage good performance. Incentive systems should be based primarily on
portfolio quality. Incentive systems that reward portfolio growth without quality will undermine
rigor in loan selection and eventually portfolio performance.

2.5. Loan Files
The loan file review will reveal whether loans are adequately documented and whether lending
operations comply with MFI credit policies. The analyst will prepare for the loan file review by
developing a procedural audit template for each type of loan. The template consists of a step-by-
step checklist of all procedures and policy guidelines associated with the analysis, approval,
conditions, disbursement and management of a loan. The analyst will typically construct this
checklist from the credit manuals and from interviews with credit department staff.

The analyst then performs the audit on a small, random sample of loan files in each branch that
reflects a range of loan products, credit loan officers, loan size, delinquency and refinance status.
These may be the same files the analyst used for the MIS audit. The selection makes no attempt
at statistical precision, but rather is an informal reflection of the composition of the branch
portfolio. The analyst runs through the checklist and fills out a template for each loan, noting
compliance or deviation from every policy and procedure. It is important to classify each
procedure so that the analyst can focus on material discrepancies and determine whether they are
clustered in a specific stage of the process (i.e., analysis, approval, disbursement, etc.). The
analysts should be especially watchful for evidence of rescheduling or refinancing of loans, or
other means of “evergreening.” 13 This should be done early in the branch office visit so the
analyst can discuss significant discrepancies with staff.

2.6. Final Analysis
Like the Tier I analysis, the Tier II analyst will be guided by judgment. However, the Tier II
exercise will provide the analyst with a wide range of opportunities to evaluate whether the MFI
conducts business according to its policies as well as to observe signs of success or failure.
Ultimately, the analyst should be able to present a well-formed opinion about the core
performance areas, substantiated by observation and analysis.

13
A loan may be rescheduled by changing the original amortization schedule. In a refinanced loan, the original loan
is “paid-off” by the new loan, which has a new amortization schedule.
Page 14
Due Diligence Guidelines for the Review of Microcredit Loan Portfolios

3. TIER 3 DUE DILIGENCE

Tier III due diligence aims to achieve a quantifiable level of certainty regarding the quality of
the loan portfolio that would satisfy a commercial investor. The exercise uses statistical methods
to verify individual account balances, measure the level of compliance with lending policies and
procedures, and check for material discrepancies and fraud.

The exercise will enable the analyst to answer three basic questions with statistical certainty:
• Do the portfolio tracking and accounting systems accurately reflect loan balances, transaction
activity, and portfolio risk?
• Does the MFI rigorously follow sound lending policies and procedures?
• Does the MFI have adequate systems of internal control to ensure robust performance
monitoring and compliance with policies and procedures?

A Tier III analysis is appropriate for a third party that requires a higher level of certainty about
portfolio performance than is normally available from an audit or ratings report. This certainty
may be necessary for commercial investors, bank regulators, or a strategic investor, any of whom
may want to establish clear benchmarks with MFI management about improvements to lending
policies and operating systems.

3.1. Audit Scope
The analyst will define the scope of the Tier III audit based on his or her assessment of the Tier
II results. Specifically, the analyst must use judgment to determine which reports, accounting
policies, and loan management procedures have significant impact on the risk of portfolio
performance, and then identify the specific data to audit. These decisions will likely require some
modifications to the scope and organization of the MIS and procedural audit templates used in
Tier II.
At a minimum, the Tier III MIS audit will go farther than the Tier II exercise by confirming the
payment vouchers and general ledger entries associated with each of the disbursements and
payments that appear in the loan tracking system.
• For each sample loan, compare the disbursements and payments that appear on the loan
summary report with the accounting vouchers that support actual disbursements and
payments. Take note especially of how payments are made (cash, check, post-dated
check, new loan, repossession of collateral, other).
• Follow each of these payments through to the general ledger, and check compliance with
the MFI’s policies for recording interest and principal amounts, and policies for non-cash
payments.
The Tier III procedural audit should include an audit of second or third loans to the same client.
The audit should confirm that the previous loans were paid off according to MFI policies.
• Determine whether the client’s prior loan was repaid by a non-cash or partial cash
payment (through rescheduling, refinancing, post-dated or third-party checks, or receipt
of collateral or equipment). This is done by reviewing the final two or three payments of

Page 15
Due Diligence Guidelines for the Review of Microcredit Loan Portfolios

the client’s prior loan to determine repayment behavior. The final payment voucher of the
prior loan should be checked to determine actual means of repayment.
• Determine whether the increase in loan and payment amounts is consistent with MFI
policy, and generally reasonable.
The analyst will also need to create a sample specifically for making field visits to the clients.
This is especially important if loan officers manage cash payments in the field. The MFI should
help introduce the analyst to the clients by presenting these visits as a “client-service evaluation.”
• The main objective of the visit is to compare the client’s records or recollection with the
repayment history and arrears status in the loan summary report. Compare their
understanding of their repayment status with the loan portfolio delinquency report.
• If a sample client is unavailable, the analyst should visit their neighborhood to ask other
clients in that vicinity to confirm the sample client’s existence and business activity.
• Determine how cash payments are actually made to the MFI. For example, are payments
made to loan officers in the field? Do group leaders make the group’s payment to the
MFI office or to a bank? Determine what documentation is provided to clients for the
cash payments. It is particularly important to focus on the time between the actual
repayment and when that repayment is entered into the MIS of the MFI.
• Determine the veracity of the basic information on the loan application and loan report.
• Look for evidence of fraud, such as loans granted to non-existent groups, kickbacks of
any of the loan amount to the originating loan officer, bribery, or other forms of
malfeasance.
• The analyst should not be accompanied by the client’s loan officer.

In the Tier III audit, the analyst will need to take special care to organize the results of the audits
so that the information can be analyzed quantitatively. Two organizational methods are
recommended. First, the results of each loan audit should be entered into a separate worksheet
template. Secondly, each template should have a summary table that quantifies the audit results.
In the MIS audit, for example, the summary table should capture the loan amount, type, actual
versus reported days in arrears, and an indicator for the validation of payment vouchers and
accounting entries. In the procedural audit, the summary table should capture the type and
number of material deviations from policy. The analyst will want to finalize the audit templates
after carefully planning which data and calculations will be required for the analysis.

3.2. The Tier III Sampling Method

The Tier III sampling methodology is designed to provide levels of certainty that are adequate
for most investors or regulators to make decisions, without incurring the enormous costs that
would be required to achieve high levels of certainty about the precise value of the portfolio. The
primary objective of this sampling method is to validate the findings of the Tier II exercise.
Specifically, the analyst will weight the sample to test the areas of high risk detected during the
Tier II analysis. This added level of statistical certainty makes Tier III conclusions uniquely
robust. The Tier III results can, for example, support the following statement:

Page 16
Due Diligence Guidelines for the Review of Microcredit Loan Portfolios

In our opinion, the MFI’s practices comply in all material respects with their policies
and procedures for loan portfolio management; internal control systems are sound,
adequate, and effective; and, accounting and portfolio reports can be relied upon to
generate accurate component balances.”

OR
In our opinion, the MFI’s practices deviate substantially from their policies and
procedures for loan portfolio management; internal control procedures are inadequate to
guarantee compliance; and, accounting and portfolio reports exaggerate actual
operating income and underreport the portfolio in arrears by as much as 20%.

Investors should be aware that the discretionary sampling in Tier III limits to some degree the
robustness of inferences about performance levels in the entire portfolio. A focused sample can
be very effective at detecting problems and estimating their dimensions, but because it is not
statistically representative of the entire population the results cannot be extrapolated to quantify a
value adjustment to the balances that appear on the MFI reports. That level of precision is
possible with larger sampling methods, but at a cost that exceeds the marginal benefits for most
investors or regulators. For most investors, the results of the discretionary sampling will be
adequate. For example, a sample of a specific loan product that constitutes 65% of the entire
portfolio could show that the MFI’s reports are inaccurate in over 20% of the loans, that over
15% loan files show material deviations from lending policies, and that the analyst detects
widespread indications of management deficiencies. In such a case, an investor or regulator has
sufficient information to take action – declining investment or launching an aggressive corrective
intervention.

3.3. Defining the Sample
The Tier III analyst makes two basic decisions about how to sample the loan portfolio. One
decision is about what part of the portfolio to sample and the other is how many loans to include
in the sample. This means that the analyst uses his or her judgment to select the types of loan to
sample, and then uses statistical models to create the sample. So for example, the analyst may
decide to sample the loans of a specific branch office, and then use statistical tables to determine
the sample size and a statistical program to randomly select the sample loans. 14

The analyst makes the first decision based on the findings of the Tier II exercise. Portfolio
performance deficiencies are typically concentrated in certain branch offices, types of loans,
credit officers, or accounting methods. A key advantage of the Tier III sampling method is
precisely that it allows the analyst to base the sample on an informed Tier II opinion about likely
problem areas. The general guiding principle is to oversample the loans that are likely to reveal
practices associated with poor performance. Typical problem areas are:
• larger loans;
• delinquent loans;
• loan types with a history of problems;
• loans to clients who were delinquent on previous loans;

14
For random selection of loans, the analyst can import a list of loans generated by the MFI’s portfolio tracking
system into EXCEL, SPSS or similar off-the-shelf software. When using EXCEL with the “Analysis Toolpak” add-
in, click on “Tools>Data Analysis>Sampling.” In SPSS, go to “Data>Select Cases” and use the option “Random.”
Page 17
Due Diligence Guidelines for the Review of Microcredit Loan Portfolios

• loans made by loan officers with higher than average loan write-offs; and,
• loans with peculiar amortization schedules that may be miscalculated in the portfolio
tracking system.

Typically, the analyst will want to design the sample so that it represents the entire portfolio as
well as focuses on probable problem areas. For example, the analyst might randomly select well
performing branch offices in addition to branch offices selected specifically for their historic
poor performance. Likewise, the global sample of loans might include a random selection from
each loan type in addition to a more robust sample of a loan type with known problems. It is very
important that the analyst justify the rationale for the discretional sampling and its extrapolation
to the portfolio.

The following is an illustrative example of Tier III sampling.

MFI Micronova has three loan products. “Business loans” were Micronova’s
first loan product and still comprise 45% of the portfolio. They are collateral-
based loans and significantly larger than any other loan product. “Microcredit”
loans are the newest and only 10% of the portfolio; they were developed with
the help of internationally experienced technical partners and are based on a
cash flow analysis of the client’s household. “Salary loans” are 30% of the
portfolio. Micronova has 12 branches. Five are at least 4 years old and
together have 70% of the entire portfolio; the other 7 branches have been in
operation less than 18 months.

The analyst is particularly concerned about the business loans, which have
higher reported arrears rates than the other loans, especially in the five oldest
and largest branches. Therefore, the analyst decides to sample 3 of the five
large branches and 2 of the 7 newer branches. Within each group the branches
are selected randomly. In each branch, the analyst decides to sample each loan
product separately. The analyst is also concerned about follow-up loans and
therefore decides to create a sample for each loan product that is half first-time
loans and half follow-up loans. Within these categories, the loans are selected
randomly.

3.4. Sample Sizes

The choice of sample size will depend upon the desired level of certainty in interpreting the
findings. 15 The relationship of sample size to confidence levels is strictly mathematical: larger
sample sizes will produce higher confidence levels and smaller margins of error in extrapolating
the sample results to the portfolio. Nevertheless, the analyst will use judgment to determine how
much certainty is required. This manual recommends a standard sizing strategy that will
probably be sufficient for most situations, but the analyst can adjust it as needed.

15
Refer to the statistical tables in the annex to determine the margins of error associated with different combinations
of sample size and proportions.
Page 18
Due Diligence Guidelines for the Review of Microcredit Loan Portfolios

The Tier III exercise uses an incremental approach to sampling that allows the analyst to stop
when the results produce the desired level of certainty. An initial sample set of 50 loans may
return results that are sufficient for some populations. For example, if a sample of 50 loans
contains 10 cases of unreported loan rescheduling, the analyst may infer, with 95% certainty, that
the incidence in the entire portfolio is between 10% and 33.7%, or a proportion of about 20%
with a margin of error of roughly 11% on either side of that value The range of this estimate is
indeed large. However, the analyst may decide that even the lower bound is unacceptable and
that the results are sufficient to confirm the initial hypothesis that serious rescheduling problems
exist. The analyst has the option of pulling another set of 50 loans for a combined sample of 100.
If the analyst were to find 22 cases in the sample of 100, the range would be 12.2% - 27.8%, or a
proportion of 22% with a lower margin of error of about 8 %.The analyst may continue to
enlarge the sample size and further reduce the margin of error. However, in many cases where
problems are detected in the first 50 loans, the slightly smaller margin of error that could be
achieved with a larger sample size will not be worth the cost.

Larger sample sizes may be desirable when the analyst wants to gain more precision for good
results in the initial sample. For example, if the analyst only finds 1 problem case in a sample of
50 loans, the 2% finding can be extrapolated to the portfolio with a margin of error that places
the incidence between 0.1% and 10.7%. A sample of 300 with 6 cases would render a smaller
margin of error with the incidence in the entire portfolio between 0.7% and 4.3%. The analyst
can use the tables in the Annex I to select the appropriate sample size.

3.5. Quantifying the Results
The final output of the audit should be a set of spreadsheets that quantify the results for each
loan. The analyst will then need to compile the results of the sample and extrapolate the results to
the portfolio. The results need to be expressed in two variables. The first is the incidence of error.
The analyst can calculate the proportion of negative findings in the sample and then consult the
tables in Annex I to determine the margin of error for extrapolating the sample findings to the
entire portfolio.

The second variable is the magnitude of the error. Here again, the analyst will have to use
judgment to determine the appropriate metric. As an example, the magnitude of discrepancies in
the arrears status of a loan should probably be measured in days (e.g., the arrears status of loans
is, on average, underreported by 17 days.) The increment in loan sizes between first, second and
third loans may be measured as a percentage of the previous loan, or in nominal amounts,
depending on how the MFI measures the increments in its own policies. In any case, when
calculating the average error, the analyst should consider the standard deviation in the data set
for clues to the underlying cause of the error. 16

The incidence and magnitude variables will enable the analyst to quantify the discrepancies. For
example, the MIS audit might show that the portfolio tracking system misstates the arrears status
in 15% of all loans and that the average error is – 34 days (arrears is typically underreported).
16
Excel (or SPSS or any statistical audit package) will calculate the median or mean average and the standard
deviation. In Excel, use the appropriate @ functions, or in the Analysis Toolpak “Tools > Data Analysis >
Descriptive Statistics”; in SPSS, use “Analyze > Descriptive Statistics > Explore.”

Page 19
Due Diligence Guidelines for the Review of Microcredit Loan Portfolios

Furthermore, a small standard deviation in the errors (meaning that the errors are fairly
consistent) might also lead the analyst to investigate whether the MFI is consistently performing
an adjustment that produces the understatement of arrears.

In the procedural audit, the results might show that 35% of loan files reflect an average of 3
material incidents of failure to comply with MFI policies. If the analyst has segmented the
procedures into operational categories (e.g., loan analysis, approval, disbursement, file
management, etc), he or she can determine whether the compliance failure is concentrated in
specific procedures or spread randomly throughout all phases of loan management.

3.6. Final Discussion with Management
At this stage of the Tier III exercise, the analyst will formulate preliminary conclusions about the
overall performance of the MFI, loan management, the integrity of policies and procedures, and
the accuracy of the portfolio reporting and accounting systems. The final step in the investigation
phase is to present these tentative conclusions to MFI management for discussion.

The final consultation with management serves at least three purposes. First, it engages
management in the final interpretation and this helps generate management buy-in to the final
report and recommendations. Secondly, it gives management an opportunity to correct
misconceptions and provide perspective that adds important nuance to the final interpretation.
And finally, the discussion gives the analyst a final look at management capacity and attitude.

3.7. Analysis and Write up
The quality of Tier III due diligence depends ultimately on the ability of the analyst to interpret
the results of the audit together with the more subjective discoveries of the Tier I and Tier II
exercises. The analyst should be able to quantify material deviations and trace them to their root
cause in deficiencies of management capacity, policies and procedures, or systems.

The Tier III report needs to be explicit about the objective of the due diligence sponsor and the
sampling method. The rationale for discretional sampling and extrapolation should be clearly
supported by the Tier I and II findings. And the precision and confidence levels of statistical
sampling should be clearly explained.

4. GENERAL COMMENTS ON PORTOFOLIO DUE DILIGENCE

In many MFIs, the due diligence exercise will reveal that actual loan portfolio performance is
weaker than it appears in official reports. The microfinance industry is still relatively young, and
few investors or regulators have developed robust systems for oversight. MFIs, just like banks,
have a strong incentive to put their best face forward when it comes to the quality of their assets.
And sometimes, their donors or investors or regulators have strong incentives for not questioning
strong performance reports. Ultimately, the analyst has to determine whether the weaknesses or
discrepancies create material risks that threaten the interests of the MFI’s donors, creditors,
investors, regulators, or depositors. The analyst’s biggest challenge is how to present the findings
in a constructive manner. The following observations are meant to provide general guidelines:

Page 20
Due Diligence Guidelines for the Review of Microcredit Loan Portfolios

• While there is a cure for ignorance, deceit is cancerous. Most MFIs are truly ignorant about
the degree or consequences of their poor MIS performance, but once trained can become
ardent disciples of transparency. In contrast, MFIs that deliberately hide portfolio
performance effectively eliminate the possibility of accurate measurement and undermine
their own credibility. In practice, an analyst can usually distinguish between those MFIs that
think transparency is good management practice and those that think transparency threatens
their access to funds.

• A weak MIS will inevitably lead to deterioration of loan portfolio performance. There should
be no doubt about this. Anytime there is a basic failure to reconcile the results from the loan
tracking system results with the accounting system, or loan delinquency reports do not reflect
transactions in a timely or accurate manner, quality will eventually degrade, no matter if
current quality looks high.

• The way that MFI management deals with a weak MIS system says much about their
commitment to accurate portfolio reporting. Ad hoc measures to monitor the portfolio may
be adequate while the MFI prepares for a long-term solution. However, hope is not a method:
faith in a future, unproven system is no substitute for a well-informed plan for creating a
robust MIS.

• There is almost never an easy fix for high levels of systemic portfolio delinquency—for
instance, when loans with payments late by more than 30 days constitute over 10% of an
microloan portfolio, or annual microloan losses are above 5% of that portfolio. Even if
current levels are not yet extreme, a trend of deteriorating repayment performance is
symptomatic of systemic problems in management, policies and procedures, and human
resources. Once a portfolio has become significantly contaminated, it takes a long time and
draconian measures to bring it back under control. These measures often include replacing
loan officers and their immediate supervisors, re-engineering the basic credit methodology,
removing substantial numbers of existing clients, writing off a large amount of unrecoverable
debt, and even changing the top management. Analysts should beware of management
promises to “get repayment levels up” over the short term if they do not have an aggressive
and concrete plan to do more than have their loan officers “work a bit harder” on this aspect
of portfolio administration.

• Organizational commitment to portfolio management starts at the top. Analysts should be
concerned about MFI managers who do not seem well informed or actively involved in
monitoring all aspects of loan recovery.

• There is often disparity between what management believes it has directed loan officers to do
and what they actually do in the field. Many top MFI managers will regard a due diligence
exercise as an opportunity to reconnect with actual practice in the field and re-establish basic
operating principles and procedures, no matter how negative the ensuing results. The analyst
should make every effort to present the findings in a way that enable management to take
constructive action.

• An astute analyst looks at trends and patterns to interpret individual findings. A problem that
is getting better is a very different matter than one that is getting worse. A due diligence
Page 21
Due Diligence Guidelines for the Review of Microcredit Loan Portfolios

sponsor certainly deserves an accurate measure of portfolio performance. Nevertheless, the
analyst’s interpretation of management commitment to performance and the trajectory of the
MFI will be the decisive factor for most investors. The analyst should be explicit about the
key factors that influence the final recommendation to an investor or regulator.

• Some potential donors or investors may want to engage in a two-step approach to the loan-
portfolio due diligence process. The first step examines practice in an initial visit, discusses
negative findings with management, and allows management a period to resolve unhealthy
practices. Then in the second step, the analyst can delve into deeper due diligence once basic
issues have been addressed. The appraisal process of can be instructive to MFI management,
showing them better ways to manage their organizations.

Page 22
Due Diligence Guidelines for the Review of Microcredit Loan Portfolios

ANNEX I: Appraisal Tables

Table 1.1
Tier I Review: MFI Operating Environment

KEY ISSUES APPRAISAL GUIDELINES
What is the growth potential of the MFI’s
• Profile the size of the target market
target market?
• Profile the lending methodology, including
loan conditions for the client (size of loan
Are the MFI’s lending products appropriate for
relative to GDP per capita and effective
the target market?
interest rate)
• Review drop out rates
• Account for other MFIs offering similar
services
• Identify competitive advantage of MFI
services vis-à-vis competitors (pay special
How is the MFI managing competition in its
attention to scope and price of services,
market?
funding and operational costs of the MFI,
and longterm funding capacity)
• Assess management strategies for dealing
with competition.
• Focus on financial performance, arrears
Do local MFIs generally comply with levels, operating cost margins, and
standards of best practice? adoption of institutional forms capable of
funding sustained growth.

Page i
Due Diligence Guidelines for the Review of Microcredit Loan Portfolios

Table 1.2
Tier I Review: Financial Reports and Accounting Policies

KEY ISSUES APPRAISAL GUIDELINES
• Anything less than publication of the full
auditor’s report on www.themix.org and
Annual financial statements are audited by a
the MFI’s website requires further
reputable firm and published in a timely
investigation into management/owner
manner.
commitment to transparency and
disclosure.
In addition to audited summary financial
• The level of detail in the financial reports
statements, the MFI has monthly financial
reflects the MFI’s ability to manage the
reports that provide adequate detail on income
loan portfolio and measure financial
sources, expenses, portfolio classification and
performance by operational departments,
provisioning, loan products, and branch
lending and funding products.
operations.
• This level of reporting will depend on the
The MFI produces daily or weekly reports for
financial structure of the MFI. The analyst
liquidity positions, asset-liability management,
should also assess whether management
funding, and any other time-sensitive
makes adequate use of these reports in the
operations.
respective areas.
• Any discrepancy in the balances of detailed
financial reports vs. audited financial
The balances of the detailed monthly reports statements is an indication of integration
reconciles precisely with the balances in the problems between components of the
audited financial statements. accounting system. Management should be
able to explain and reconcile any
differences.
• Any discrepancies here warrant careful
The portfolio balances in the loan tracking scrutiny to ensure consistent and verifiable
reports reconcile precisely with the balances accounting of loan performance from the
that appear on the balance sheet. loan registers all the way to the balance
sheet.
• PAR reports are necessary for an accurate
accounting and management of loans in
The portfolio report classifies the outstanding arrears.
balance of every loan by arrears categories • The arrears classification categories and
(this is a portfolio-at-risk classification of the corresponding loss provisions must be
entire portfolio). consistent with international standards or
the MFI’s documented historical
experience.
• These loans should be classified separately
Refinanced or rescheduled loans are clearly
and subject to specific provisioning
segmented in the portfolio classification.
policies.

Page ii
Due Diligence Guidelines for the Review of Microcredit Loan Portfolios

KEY ISSUES APPRAISAL GUIDELINES
• These policies and procedures should be
clearly documented and their application
Accounting for loan loss provisions, provision readily verified in the financial statements.
expenses, write-offs, write-off recoveries, and A comprehensive set of policies will enable
interest accrual on non-performing loans are the analyst to calculate all key indicators
clearly presented in the financial statements. related to PAR and loan provisioning/loss
expenses over several accounting periods.
Policies that obscure clear accounting in
this area are a sign of inadequate portfolio
management.
• A robust accounting system will be
What are the accounting entries related to an reflected in portfolio reports that reconcile
individual loan? Are disbursement, interest individual account detail with the global
accrual and repayment transactions precise amounts that appear on the financial
enough to ensure accurate individual loan statements. Anything less is a limit on
tracking and aggregate portfolio reporting? management’s ability to supervise the
performance of loan officers and their
individual loans.
• Policies governing reversal of accrued
Is accrued interest income reversed when a income should be consistent with the
loan falls into arrears? How is interest income nature and tenor of the loan. As a general
accrued? What is the treatment of accruals? rule, accrued interest should be reversed
after no more than 90 days of arrears.
• Is unrecorded cash in the hands of loan
officers for any amount of time between
How is cash related to loans paid out, received, the point of payment and recording in the
and controlled? loan tracking system? (Controls should
exist to ensure that cash is accounted for
immediately and accurately.)
• Widespread practice of booking loan
Can loans be paid off through “in-kind”
repayments before the cash is received can
payments, receipt of collateral guarantees,
significantly exaggerate actual income. The
post-dated checks, rescheduled loans, or any
accounting of these transactions should
other non-cash means?
provide a clear trail to the actual cash
receipt or write-down.

Page iii
Due Diligence Guidelines for the Review of Microcredit Loan Portfolios

Table 1.3

Tier I Review: Financial Performance
KEY ISSUES APPRAISAL GUIDELINES
Key performance indicators are:
• Portfolio yield
• PAR
Overall financial performance is consistent
• Provisioning and loan loss expenses
with industry standards for similar MFIs
• Operating cost ratio, or cost per borrower
as a percent of per capita GDP
• ROA
• Persistent arrears are a sign of management
complacency and poor credit culture.
• Portfolio trends should be consistent. The
analyst should watch for signs of abrupt
Portfolio quality is good over time, as changes in performance.
measured by portfolio at risk (PAR). • As a general rule, MFIs that have total
loans more than 30 days in arrears that are
greater than 10% of the portfolio, or annual
loan losses greater than 5%, are unlikely to
maintain delinquency at sustainable levels.
• Yield gap analysis is a quick check on
The effective portfolio yield is consistent with
hidden non-performing loans in the
expected yields.
portfolio.
Rapid growth inevitably leads to stress on
systems and collection performance. Growth
stress is typically evident in these indicators:
• Portfolio quality
Is the institution growing at a sustainable rate? • Return on portfolio
• # of clients or disbursements
• # of clients per credit officer
• Staff turnover
• Client turnover
• Significant restriction in credit
Is the MFI facing liquidity constraints that
disbursements breaks implicit contracts and
affect its ability to disburse loans?
is usually associated with deterioration of
credit discipline.

Page iv
Due Diligence Guidelines for the Review of Microcredit Loan Portfolios

Table 1.4
Tier I Review: Credit Policies and Procedures

KEY ISSUES APPRAISAL GUIDELINES
Are loan appraisal and approval criteria and
• Guidelines may need to be specific to
procedures clearly documented and adequate to
different loan products.
ensure consistent loan analysis?
• Forms should be comprehensive and
Do the application and approval forms reflect
organized to facilitate efficient loan
the loan analysis and approval procedures?
processing.
• The analyst should be able to diagram the
different levels of approval authority and
readily identify the control mechanisms.
Is the approval process structured with
• Authority levels should be appropriate to
adequate levels of authority and control?
loan amounts and related risk exposure.
• The system should reinforce the
accountability of the credit officer who
recommends loans for approval.
• The definition should include all loans that
have had their original terms altered either
How are loan restructuring and renegotiation in amount, interest rate, or amortization
defined and treated? period.
• Check to see if the MFI allows new loans
to a client whose prior delinquent loans
have not been paid off.
What is the institution’s policy regarding
• Multiple lending that is not clearly reported
multiple, concurrent loans to an individual
to show the total borrowing by an
client?
individual client is a recipe for high default.

Does the MFI have clear policies about the size • Lack of analysis of client repayment
of loans and loan payments relative to the the capacity will typically lead to eventual
client’s ability to pay? repayment problems.
What is the rate of increase in loan amounts • Increases should be incremental. The
and average size of payments for repeat history of arrears will indicate whether the
borrowers? increments are appropriate.
• Obligatory savings reduce the effective
How big are any obligatory savings
loan amount and at some level lead to high
requirements relative to loan amounts?
client turnover.

Page v
Due Diligence Guidelines for the Review of Microcredit Loan Portfolios

Table 1.5
Tier I Review: MIS and Performance Reporting

KEY ISSUES APPRAISAL GUIDELINES
• The MIS should be able to generate a client
report with a complete history of loans and
transactions.
• Credit officers should have current reports
Does the MIS generate a complete set of on applicants and clients.
reports that cover all necessary areas of credit • Credit supervisors should have reports on
operations? the activities of each credit officer.
• Credit managers should have reports by
branch and lending products.
• Senior Management and the Board of
Directors should have summary reports.
• The analyst should be alert to any
comments from the operations, accounting
Are the reports timely and accurate? or internal control staff that the credit
reports are outdated or not accurate. Claims
of inaccuracy should be followed up.
Can the MIS produce reports to show trends • This is critical for trend analysis and
over time? strategic decision making.
• Access policies and procedures should
facilitate efficient data entry and report
generation while guarding against fraud or
Is access to the MIS strictly controlled by a
data manipulation.
hierarchy of access privileges.
• Management and staff attitudes about
control provide important clues to the
integrity of the system.
Does the IT department follow a strict protocol • Infrastructure security must be robust
for back up and safeguarding of electronic enough to safeguard records in case of
records? catastrophe.
• MIS audits should be included in annual
Does the MFI ask its external auditors to carry audits as a fraud control measure.
out MIS audits? • Check the terms of reference for the last
audit.
Is the IT infrastructure and MIS software • The analyst should consider whether the
adequate for the MFI’s next phase of growth? MFI has the capacity to expand its system.

Page vi
Due Diligence Guidelines for the Review of Microcredit Loan Portfolios

Table 1.6
Tier I Review: Portfolio Monitoring and Control

KEY ISSUES APPRAISAL GUIDELINES
• This will be reflected in widespread
Does senior management set targets and understanding of the targets, frequent
monitor loan delinquency on a routine basis? delinquency reports, and well-defined
action plans for the different levels of staff
involved in credit operations.
• The analyst should make the subjective
assessment of whether management’s
Can senior management clearly explain the
understanding of loan performance is
causes and cycles of arrears?
supported by portfolio reports and the
views of operational staff.
• It is important that management and staff
can track portfolio performance by these
Do the portfolio reports show performance by key factors. Without this breakdown,
loan product, branch, and credit officer? management is not able to identify the
sources of strength or weakness in credit
operations.
Does the institution track payment • This capacity will also be reflected in loan
performance in ways that help it understand appraisal methods and lending products
client behavior? that are specific to client groups.
• Staff who are not held accountable will not
produce optimal results over time.
Do staff incentives include loan portfolio
quality as a primary performance marker. • Incentive systems that reward growth
without quality will encourage poor loan
selection and lead to poor performance.
What are the ex-post internal controls
employed by the institution, independent of the
• In a maturing MFI, internal control systems
operations department, that review loan
must be robust and well supported by
approvals, scrutinize problem credits, evaluate
management and the board of directors.
provision coverage, look at trends, and cite
documentation exceptions?
• The analyst needs to determine if the credit
Who handles collection of late payments, the
officer has sufficient incentive to ensure
original loan officers or a specialist?
that his or her loans are performing.
At what point is a loan sent for collection • The legal remedy must be efficient and
through the legal system? How effective is this cost-effective. If not, problems will likely
remedy and how long does it usually take to accumulate over time and result in large
collect? periodic write-offs.
• Fraud control and remedies should be
Has fraud played a role in losses?
robust. The analyst should ask about past
cases.

Page vii
Due Diligence Guidelines for the Review of Microcredit Loan Portfolios

Table 2.1
Tier II Review: MIS Review

Note: This table guides the comparison of the results from the analysts’ MIS spreadsheet
template and the MIS loan report.

KEY ISSUES APPRAISAL GUIDELINES
• This is primarily to a test to see if the MIS
Does the MIS accurately classify individual
internal calculations on arrears are
loan principal balances by arrears status?
accurate.
• Check to see that the classification of
Does the MIS accurately assign individual
individual loans is consistent with MFI
loans to the correct provisioning category
policy.
Do individual MIS loan accounts accurately Using the spreadsheet template, the analyst
reflect the contractual conditions of the loan will verify consistency in the following loan
contract? conditions:
• loan amortization
Do individual MIS loan accounts accurately • interest calculation
reflect MFI lending guidelines and accounting • fees
policies? • distribution of payments between principal,
interest and fees
Does the system generate a schedule for each • These reports are necessary for tracking
client that compares actual payment dates and client and portfolio performance against the
amounts to the original amortization schedule? expected repayment schedules.
And does the system generate a consolidated • This information is necessary for
actual vs scheduled repayment report? performing the yield gap analysis.
• This comparison requires a confirmation of
Are actual disbursements and payments
the dates on the physical payment and
accurately entered into the MIS?
disbursement vouchers.

Page viii
Due Diligence Guidelines for the Review of Microcredit Loan Portfolios

Table 2.2
Tier II Review: Loan Classification

KEY ISSUES APPRAISAL GUIDELINES
Does the MFI have clearly defined criteria for • Some MFIs may consider other factors in
assigning loans to risk categories? addition to arrears.
• The analyst can test for this on the small
Are risk categories applied consistently to
Tier II sample of loans; robust assessment
individual loans?
will require Tier III analysis.
• This will involve calculating write-offs as a
percentage of portfolio and tracking the
evolution of risk category balances over
Are provisioning levels consistent with historic time.
write-off activity • If the MFI has not done such an analysis,
are provisioning levels appropriately
conservative in relation to international
benchmarks?

Page ix
Due Diligence Guidelines for the Review of Microcredit Loan Portfolios

Table 2.3
Tier II Review: Branch Office Management

KEY ISSUES APPRAISAL GUIDELINES
Specific areas of importance:
Does the distribution of decision making • determination of lending targets;
authority and operational responsibility • loan approvals;
between headquarters and the branch offices • disbursement and payment cash
facilitate efficient and controlled operations? management;
• data entry into loan tracking system;
• accounting reconciliation.
Specific areas of importance:
• determination of lending and delinquency
targets;
Does the MFI have a robust system for
• performance data collection;
managing the performance of branch office
• performance data reporting;
operations?
• performance review;
• follow up actions for good and poor
performance
The analyst should assess management and
Does the MFI adequately manage the core staff understanding and commitment to:
functions related to portfolio quality? • loan analysis; and
• delinquency management
• Check for discrepancies or disagreements
Are branch office practices consistent with around policies;
headquarters policy and expectations? • Are there significant differences between
branch offices?
• Check branch office staff attitudes about
headquarter policies and management;
Is staff morale positive?
• Note level of staff enthusiasm for reaching
performance targets

Page x
Due Diligence Guidelines for the Review of Microcredit Loan Portfolios

Table 2.4
Tier II Review: Loan Officers

KEY ISSUES APPRAISAL GUIDELINES
• What is the typical education and
employment background of a credit
Are loan officers adequately trained?
officer?
• Are the MFI training programs adequate?
Do loan officers have an accurate • Lack of clarity on lending products should
understanding of policies related to each alert the analyst to the possibility of
lending product? widespread problems in the portfolio.
• This can be observed in discussions about
Do loan officers speak knowledgably about the
how the credit officer uses personal
credit risk of their clientele and how to
judgment during loan analysis and dealing
evaluate and control those risks?
with delinquent clients.
• Report frequency should be daily or
Do loan officers receive frequent reports on weekly to keep the credit officer up to date
their portfolio performance? on pending loan applications, arrears status,
and priority cases.
• Credit officers should be very clear about
procedures and remedies and show
Do loan officers follow vigorous procedures dedication to their implementation.
for dealing with delinquent loans? • Policies should specify how quickly a
delinquent borrower must be contacted, and
by whom.
• In general, credit officers should be
spending the majority of their time with
How often do credit officers see clients, and is
clients at their place of business, or at their
this in the office or at their place of business?
group meeting site. Any deviation from this
norm should be well justified.

Page xi
Due Diligence Guidelines for the Review of Microcredit Loan Portfolios

Table 2.
Tier II Review: Loan Files

Note: This table provides general guidelines for interpreting the results of the procedural audit on
the loan files.

KEY ISSUES APPRAISAL GUIDELINES
Does the MFI’s loan file template adequately • This is ultimately a judgment about
document the essential components of loan whether the files have the appropriate level
appraisal, approval and management? of documentation.
• This is a question of whether the loan files
actually contain the documents they are
Are the loan files complete?
supposed to. A complete assessment is only
possible in Tier III.
• Typically, the level of compliance will be
consistent with the analyst’s other findings
Do deviations from policy or incomplete files
about how well the MFI conducts business
constitute a material deficiency?
according to well defined policies and
procedures.

Page xii
Due Diligence Guidelines for the Review of Microcredit Loan Portfolios

TABLE FOR EXTRAPOLATING SAMPLE RESULTS TO TOTAL PORTFOLIO
AT 95% CONFIDENCE LEVEL

SAMPLE SIZE
25 50 75 100
CASES lower P upper lower P upper lower P upper lower P upper
0 0.0% 0.0% 12.0% 0.0% 0.0% 6.0% 0.0% 0.0% 4.0% 0.0% 0.0% 3.0%
1 0.1% 4.0% 20.4% 0.1% 2.0% 10.7% 0.0% 1.3% 7.2% 0.0% 1.0% 5.5%
2 1.0% 8.0% 26.0% 0.5% 4.0% 13.7% 0.3% 2.7% 9.3% 0.2% 2.0% 7.0%
3 2.6% 12.0% 31.2% 1.3% 6.0% 16.6% 0.8% 4.0% 11.3% 0.6% 3.0% 8.5%
4 4.5% 16.0% 36.1% 2.2% 8.0% 19.2% 1.5% 5.3% 13.1% 1.1% 4.0% 9.9%
5 6.8% 20.0% 40.7% 3.3% 10.0% 21.8% 2.2% 6.7% 14.9% 1.6% 5.0% 11.3%
6 9.4% 24.0% 45.1% 4.5% 12.0% 24.3% 3.0% 8.0% 16.6% 2.2% 6.0% 12.6%
7 12.1% 28.0% 49.4% 5.8% 14.0% 26.7% 3.8% 9.3% 18.3% 2.9% 7.0% 13.9%
8 15.0% 32.0% 53.5% 7.2% 16.0% 29.1% 4.7% 10.7% 19.9% 3.5% 8.0% 15.2%
9 18.0% 36.0% 57.5% 8.6% 18.0% 31.4% 5.6% 12.0% 21.6% 4.2% 9.0% 16.4%
10 21.1% 40.0% 61.3% 10.0% 20.0% 33.7% 6.6% 13.3% 23.2% 4.9% 10.0% 17.6%
11 24.5% 44.0% 63.5% 10.5% 22.0% 33.5% 6.7% 14.7% 22.7% 4.9% 11.0% 17.1%
12 28.4% 48.0% 67.6% 12.2% 24.0% 35.8% 7.7% 16.0% 24.3% 5.6% 12.0% 18.4%
13 32.4% 52.0% 71.6% 13.8% 26.0% 38.2% 8.8% 17.3% 25.9% 6.4% 13.0% 19.6%
14 36.5% 56.0% 75.5% 15.6% 28.0% 40.4% 9.8% 18.7% 27.5% 7.2% 14.0% 20.8%
15 38.7% 60.0% 78.9% 17.3% 30.0% 42.7% 10.9% 20.0% 29.1% 8.0% 15.0% 22.0%
16 42.5% 64.0% 82.0% 19.1% 32.0% 44.9% 12.1% 21.3% 30.6% 8.8% 16.0% 23.2%
17 46.5% 68.0% 85.1% 20.9% 34.0% 47.1% 13.2% 22.7% 32.1% 9.6% 17.0% 24.4%
18 50.6% 72.0% 87.9% 22.7% 36.0% 49.3% 14.3% 24.0% 33.7% 10.5% 18.0% 25.5%
19 54.9% 76.0% 90.6% 24.5% 38.0% 51.5% 15.5% 25.3% 35.2% 11.3% 19.0% 26.7%
20 59.3% 80.0% 93.2% 26.4% 40.0% 53.6% 16.7% 26.7% 36.7% 12.2% 20.0% 27.8%
21 63.9% 84.0% 95.5% 28.3% 42.0% 55.7% 17.8% 28.0% 38.2% 13.0% 21.0% 29.0%
22 68.8% 88.0% 97.5% 30.2% 44.0% 57.8% 19.0% 29.3% 39.6% 13.9% 22.0% 30.1%
23 74.0% 92.0% 99.0% 32.2% 46.0% 59.8% 20.2% 30.7% 41.1% 14.8% 23.0% 31.2%
24 79.7% 96.0% 99.9% 34.2% 48.0% 61.8% 21.4% 32.0% 42.6% 15.6% 24.0% 32.4%
25 88.0% 100.0% 100.0% 36.1% 50.0% 63.9% 22.7% 33.3% 44.0% 16.5% 25.0% 33.5%
26 38.2% 52.0% 65.8% 23.9% 34.7% 45.4% 17.4% 26.0% 34.6%
27 40.2% 54.0% 67.8% 25.1% 36.0% 46.9% 18.3% 27.0% 35.7%
28 42.2% 56.0% 69.8% 26.4% 37.3% 48.3% 19.2% 28.0% 36.8%
29 44.3% 58.0% 71.7% 27.6% 38.7% 49.7% 20.1% 29.0% 37.9%
30 46.4% 60.0% 73.6% 28.9% 40.0% 51.1% 21.0% 30.0% 39.0%
31 48.5% 62.0% 75.5% 30.2% 41.3% 52.5% 21.9% 31.0% 40.1%
32 50.7% 64.0% 77.3% 31.5% 42.7% 53.9% 22.9% 32.0% 41.1%
33 52.9% 66.0% 79.1% 32.8% 44.0% 55.2% 23.8% 33.0% 42.2%
34 55.1% 68.0% 80.9% 34.1% 45.3% 56.6% 24.7% 34.0% 43.3%
35 57.3% 70.0% 82.7% 35.4% 46.7% 58.0% 25.7% 35.0% 44.3%
36 59.6% 72.0% 84.4% 36.7% 48.0% 59.3% 26.6% 36.0% 45.4%
37 61.8% 74.0% 86.2% 38.0% 49.3% 60.6% 27.5% 37.0% 46.5%
38 64.2% 76.0% 87.8% 39.4% 50.7% 62.0% 28.5% 38.0% 47.5%
39 66.5% 78.0% 89.5% 40.7% 52.0% 63.3% 29.4% 39.0% 48.6%
40 66.3% 80.0% 90.0% 42.0% 53.3% 64.6% 30.4% 40.0% 49.6%
41 68.6% 82.0% 91.4% 43.4% 54.7% 65.9% 31.4% 41.0% 50.6%
42 70.9% 84.0% 92.8% 44.8% 56.0% 67.2% 32.3% 42.0% 51.7%
43 73.3% 86.0% 94.2% 46.1% 57.3% 68.5% 33.3% 43.0% 52.7%
44 75.7% 88.0% 95.5% 47.5% 58.7% 69.8% 34.3% 44.0% 53.7%
45 78.2% 90.0% 96.7% 48.9% 60.0% 71.1% 35.2% 45.0% 54.8%
46 80.8% 92.0% 97.8% 50.3% 61.3% 72.4% 36.2% 46.0% 55.8%
47 83.5% 94.0% 98.8% 51.7% 62.7% 73.6% 37.2% 47.0% 56.8%
48 86.3% 96.0% 99.5% 53.1% 64.0% 74.9% 38.2% 48.0% 57.8%
49 89.4% 98.0% 100.0% 54.6% 65.3% 76.1% 39.2% 49.0% 58.8%
50 94.0% 100.0% 100.0% 56.0% 66.7% 77.3% 40.2% 50.0% 59.8%

Page xiii
Due Diligence Guidelines for the Review of Microcredit Loan Portfolios

TABLE FOR EXTRAPOLATING SAMPLE RESULTS TO TOTAL PORTFOLIO
AT 95% CONFIDENCE LEVEL (continued)

SAMPLE SIZE
150 200 250 300
CASES lower P upper lower P upper lower P upper lower P upper
0 0.0% 0.0% 2.0% 0.0% 0.0% 1.5% 0.0% 0.0% 1.2% 0.0% 0.0% 1.0%
1 0.0% 0.7% 3.7% 0.0% 0.5% 2.8% 0.0% 0.4% 2.2% 0.0% 0.3% 1.8%
2 0.2% 1.3% 4.7% 0.1% 1.0% 3.6% 0.1% 0.8% 2.9% 0.1% 0.7% 2.4%
3 0.4% 2.0% 5.7% 0.3% 1.5% 4.3% 0.3% 1.2% 3.5% 0.2% 1.0% 2.9%
4 0.7% 2.7% 6.7% 0.6% 2.0% 5.0% 0.4% 1.6% 4.1% 0.4% 1.3% 3.4%
5 1.1% 3.3% 7.6% 0.8% 2.5% 5.7% 0.7% 2.0% 4.6% 0.5% 1.7% 3.9%
6 1.5% 4.0% 8.5% 1.1% 3.0% 6.4% 0.9% 2.4% 5.2% 0.7% 2.0% 4.3%
7 1.9% 4.7% 9.4% 1.4% 3.5% 7.1% 1.1% 2.8% 5.7% 0.9% 2.3% 4.8%
8 2.3% 5.3% 10.2% 1.7% 4.0% 7.7% 1.4% 3.2% 6.2% 1.2% 2.7% 5.2%
9 2.8% 6.0% 11.1% 2.1% 4.5% 8.4% 1.7% 3.6% 6.7% 1.4% 3.0% 5.6%
10 3.2% 6.7% 11.9% 2.4% 5.0% 9.0% 1.9% 4.0% 7.2% 1.6% 3.3% 6.0%
11 3.2% 7.3% 11.5% 2.3% 5.5% 8.7% 1.9% 4.4% 6.9% 1.5% 3.7% 5.8%
12 3.7% 8.0% 12.3% 2.7% 6.0% 9.3% 2.2% 4.8% 7.4% 1.8% 4.0% 6.2%
13 4.2% 8.7% 13.2% 3.1% 6.5% 9.9% 2.4% 5.2% 8.0% 2.0% 4.3% 6.6%
14 4.7% 9.3% 14.0% 3.5% 7.0% 10.5% 2.7% 5.6% 8.5% 2.3% 4.7% 7.1%
15 5.2% 10.0% 14.8% 3.8% 7.5% 11.2% 3.1% 6.0% 8.9% 2.5% 5.0% 7.5%
16 5.7% 10.7% 15.6% 4.2% 8.0% 11.8% 3.4% 6.4% 9.4% 2.8% 5.3% 7.9%
17 6.3% 11.3% 16.4% 4.6% 8.5% 12.4% 3.7% 6.8% 9.9% 3.1% 5.7% 8.3%
18 6.8% 12.0% 17.2% 5.0% 9.0% 13.0% 4.0% 7.2% 10.4% 3.3% 6.0% 8.7%
19 7.3% 12.7% 18.0% 5.4% 9.5% 13.6% 4.3% 7.6% 10.9% 3.6% 6.3% 9.1%
20 7.9% 13.3% 18.8% 5.8% 10.0% 14.2% 4.6% 8.0% 11.4% 3.8% 6.7% 9.5%
21 8.4% 14.0% 19.6% 6.3% 10.5% 14.7% 5.0% 8.4% 11.8% 4.1% 7.0% 9.9%
22 9.0% 14.7% 20.3% 6.7% 11.0% 15.3% 5.3% 8.8% 12.3% 4.4% 7.3% 10.3%
23 9.6% 15.3% 21.1% 7.1% 11.5% 15.9% 5.6% 9.2% 12.8% 4.7% 7.7% 10.7%
24 10.1% 16.0% 21.9% 7.5% 12.0% 16.5% 5.9% 9.6% 13.3% 4.9% 8.0% 11.1%
25 10.7% 16.7% 22.6% 7.9% 12.5% 17.1% 6.3% 10.0% 13.7% 5.2% 8.3% 11.5%
26 11.3% 17.3% 23.4% 8.3% 13.0% 17.7% 6.6% 10.4% 14.2% 5.5% 8.7% 11.9%
27 11.9% 18.0% 24.1% 8.8% 13.5% 18.2% 7.0% 10.8% 14.6% 5.8% 9.0% 12.2%
28 12.4% 18.7% 24.9% 9.2% 14.0% 18.8% 7.3% 11.2% 15.1% 6.0% 9.3% 12.6%
29 13.0% 19.3% 25.7% 9.6% 14.5% 19.4% 7.6% 11.6% 15.6% 6.3% 9.7% 13.0%
30 13.6% 20.0% 26.4% 10.1% 15.0% 19.9% 8.0% 12.0% 16.0% 6.6% 10.0% 13.4%
31 14.2% 20.7% 27.1% 10.5% 15.5% 20.5% 8.3% 12.4% 16.5% 6.9% 10.3% 13.8%
32 14.8% 21.3% 27.9% 10.9% 16.0% 21.1% 8.7% 12.8% 16.9% 7.2% 10.7% 14.2%
33 15.4% 22.0% 28.6% 11.4% 16.5% 21.6% 9.0% 13.2% 17.4% 7.5% 11.0% 14.5%
34 16.0% 22.7% 29.4% 11.8% 17.0% 22.2% 9.4% 13.6% 17.8% 7.7% 11.3% 14.9%
35 16.6% 23.3% 30.1% 12.2% 17.5% 22.8% 9.7% 14.0% 18.3% 8.0% 11.7% 15.3%
36 17.2% 24.0% 30.8% 12.7% 18.0% 23.3% 10.0% 14.4% 18.8% 8.3% 12.0% 15.7%
37 17.8% 24.7% 31.6% 13.1% 18.5% 23.9% 10.4% 14.8% 19.2% 8.6% 12.3% 16.1%
38 18.4% 25.3% 32.3% 13.6% 19.0% 24.4% 10.7% 15.2% 19.7% 8.9% 12.7% 16.4%
39 19.0% 26.0% 33.0% 14.0% 19.5% 25.0% 11.1% 15.6% 20.1% 9.2% 13.0% 16.8%
40 19.6% 26.7% 33.7% 14.5% 20.0% 25.5% 11.5% 16.0% 20.5% 9.5% 13.3% 17.2%
41 20.2% 27.3% 34.5% 14.9% 20.5% 26.1% 11.8% 16.4% 21.0% 9.8% 13.7% 17.6%
42 20.8% 28.0% 35.2% 15.4% 21.0% 26.6% 12.2% 16.8% 21.4% 10.1% 14.0% 17.9%
43 21.4% 28.7% 35.9% 15.8% 21.5% 27.2% 12.5% 17.2% 21.9% 10.4% 14.3% 18.3%
44 22.0% 29.3% 36.6% 16.3% 22.0% 27.7% 12.9% 17.6% 22.3% 10.7% 14.7% 18.7%
45 22.7% 30.0% 37.3% 16.7% 22.5% 28.3% 13.2% 18.0% 22.8% 11.0% 15.0% 19.0%
46 23.3% 30.7% 38.0% 17.2% 23.0% 28.8% 13.6% 18.4% 23.2% 11.3% 15.3% 19.4%
47 23.9% 31.3% 38.8% 17.6% 23.5% 29.4% 14.0% 18.8% 23.6% 11.6% 15.7% 19.8%
48 24.5% 32.0% 39.5% 18.1% 24.0% 29.9% 14.3% 19.2% 24.1% 11.9% 16.0% 20.1%
49 25.2% 32.7% 40.2% 18.5% 24.5% 30.5% 14.7% 19.6% 24.5% 12.2% 16.3% 20.5%
50 25.8% 33.3% 40.9% 19.0% 25.0% 31.0% 15.0% 20.0% 25.0% 12.4% 16.7% 20.9%

Page xiv
Due Diligence Guidelines for the Review of Microcredit Loan Portfolios

NOTES ON STATISTICAL TABLES

The statistical tables show that for each sample size, the proportion of “cases” of error or of other
occurrences can be extrapolated to the entire portfolio within a specific margin of error. There
are three columns in the table for each sample size. The middle (shaded) column shows the
proportion of the sample represented by the number of cases. The left column shows the lower
bound of what the proportion may actually be in the total portfolio; the right column shows the
upper bound (both of these are at a 95% confidence level).

The following example illustrates how to use the table:
• an analyst draws a sample of 75 loans and finds 6 cases of error (for example, the arrears
classification is incorrect);
• where the row that begins with 6 cases intersects with the middle column in the 75 Sample
Size section, the table shows the proportion of 8%;
• the left column of the section shows a lower bound of 3.0% and the right column an upper
bound of 16.6%.
• this means the following: The sample tests indicate that the most probable rate of incorrect
arrears classification in the portfolio is 8%. The margin of error for the estimate indicates
that the true proportion of the portfolio may be as low as 3% or as high as 16.6%, with a
95% confidence level—that is to say, there is only a one in twenty chance that the true
proportion is below 3% or above 16.6%.

The analyst may also be able to use the underlying formulas to calculate proportions and margins
of error for sample sizes that do not appear in the chart. The following statistical conventions
apply:

• Most of the table is derived from the normal approximation to the distribution of proportions,
using a 95% confidence interval:

Where:
πL= lower bound of distribution in the population
πU= upper bound of distribution in the population
p = proportion of the sample
n = sample size

Example: In a sample of 75 with 11 cases, the margin of error would be:

.1467(1 − .1467)
1.96 = .08
75

Therefore, the proportion of the population is 14.67% (11/75), with an 8% margin of error. This
means that the actual proportion is between 6.67% and 22.67%, with 95% confidence .

Page xv
Due Diligence Guidelines for the Review of Microcredit Loan Portfolios

• However, a more precise formula 17 based directed on the binomial distribution is used where
the accuracy of the normal approximation breaks down 18, which is where:

np (1 − p ) ≤ 8

For example, in a sample of 40 with 3 cases, the approximation formula is not accurate enough
to calculate the margins of error because 40*.075*(1-.075)= 2.775, which is well below 8. In this
case, the analyst is advised to use one of the sample sizes in the table or consult a statistician
capable of running the precise formula. In the tables above, all sample results with 10 or less
cases are calculated with the precise formula.

• Finally, for samples of any size with 0 cases, the table uses the rule of thumb that the 95%
confidence interval runs from 0 to 3/n. 19 For example, in a sample of 50 with 0 cases, the
proportion in the population is between 0% and 6%.

17
The formula is implemented as an iterative process. See Clopper, C.J., and E.S. Pearson. 1934. Use of confidence
or fiducial limits illustrated in the case of the binomial. Biometrika 26: 404-413.
18
Van Belle, G., L. Fisher, P.J. Heagerty, and T. Lumley. 2004. Biostatistics: A Methodology for the Health
Sciences, 2nd edition. Wiley. Note: the authors actually recommend using the precise formula when result of the
equation is less than 10. However, the precision above 8 is adequate for auditing purposes.
19
Joseph L. and Reinhold C., Statistical Inference for Proportions. AJR 2005;184: 1057.

Page xvi
Due Diligence Guidelines for the Review of Microcredit Loan Portfolios

ANNEX 2: Yield Gap Analysis

Yield gap analysis determines how the cash that the MFI is collecting on its loans (i.e., actual
yield) compares with the cash the MFI should be collecting if all the loan contracts are honored
in full (i.e., expected yield). The basic formula is

Cash received during the period
1 - ________________________________
Cash due during the period

It is usually best to include only cash receipts when computing the actual yield. Non-cash loan
repayments such as seizure of collateral, acceptance of post-dated checks, or retirement of the
old loan by executing a new one, often simply defer recognition of losses. For purposes of the
yield gap computation, such non-cash payments are recognized only if and when they are
converted into cash (e.g., when collateral is actually sold).

Depending on the information available, calculating a yield gap can be more or less complicated.
This Annex does not try to provide detailed instructions; rather, it presents a general approach
that the analyst may need to fine-tune to fit individual circumstances.

The analysis is normally conducted on interest and fee income, or on interest income only if all
fees are collected or netted out at the time the loan is disbursed. Late payment penalties (as
opposed to late payment of scheduled interest) are normally excluded from the calculation. In
exceptional cases where the MFI’s accounting system cannot reliably separate interest from
principal in individual loan payments, the analysis can be done using total payments rather than
just the portion attributable to interest income.

The main challenge is calculating the denominator of the fraction, cash due. The best way to
calculate this expected yield for a period is to cumulate the payments that fall due for the first
time during that period, in accordance with the original amortization schedules of the individual
loans. The concept here is that the MIS maintains a register with a series of “buckets,” one for
each month, to record expected payments that are due during that month. At the time a loan is
issued, each of its expected payments is added to the appropriate bucket. If the borrower later
misses a payment, that payment obligation stays in its original cash-due bucket, rather than being
moved to a later one. Likewise, when a loan is pre-paid, the original amortization schedule as
distributed among the buckets is not changed. This treatment of late payments and prepayments
prevents double-counting. However, it will make the ratio of cash paid to cash due jump around
from month to month, rising above 100% in months when unusual amounts of late payments or
prepayments are received. Thus, the yield gap analysis using this method would normally be
done for a full year, to even out random fluctuation and seasonal variations.

This analysis is similar to the calculation of a “current repayment rate” as discussed in
Rosenberg, Measuring Microcredit Delinquency, CGAP Occasional Paper No. 3, June 1999, pp.
8-13. Treatment of renegotiated loans is discussed there.

Page i
Due Diligence Guidelines for the Review of Microcredit Loan Portfolios

If the MIS cannot produce the above-described register of expected payments (i.e., a cumulative
amortization schedule), a less precise alternative is to generate an expected yield by analyzing
the effective yield of the MFI’s loan contracts. 20 This is easiest when all of the MFI’s loans have
the same effective interest yield. Where there are multiple products and multiple yields, the
analyst needs to multiply the yield for each product times the average outstanding balance for
that product over the period. Note that annual average balance (computed by adding the balances
at the beginning and end of the year, and dividing by two) is not precise enough to be useful. At
a minimum one would need a monthly average balance for the year (the sum of the balances at
the beginning of the year and at the end of each month, divided by 13). This method of
estimating the expected yield needs to be implemented with great care, including assurance that
the actual recognition of interest income when it is booked into the MFI’s accounting system
corresponds to the way interest is treated in calculating the expected yield.

An alternative to the above methods is to (1) draw a random sample from each category of loan,
(2) calculate repayment rates for each category by comparing actual cash received during the
year with the cash that was due under the original loan terms, (3) subtract the collection rates
from 1.00 to generate a yield gap for each category in the sample, (4) multiply these yield gaps
by the average annual outstanding balance of the corresponding part of the total portfolio, and (5)
divide the result by the annual average balance of the total portfolio.

Once the yield gap is calculated, the final step is to determine whether it is in line with the level
of delinquency/default that the MFI is reporting. 21 If the yield gap is substantially larger than
what it should be, the explanation must be determined. The commonest explanations in order of
frequency are unrecognized delinquency/default, accounting problems, and fraud.

20
Calculation of an annual “effective interest rate” normally includes two operations: (1) determining what the
interest rate for a single period would be if calculated on a declining-balance basis, and (2) compounding that
interest rate over the number of periods in the year. For yield gap purposes, one would calculate an “effective yield,”
by which we refer to an uncompounded annual rate: i.e., the period rate determined in step (1) is simply multiplied
by the periods per year. If the declining balance rate is 2 percent per month, the annual effective interest rate would
be 1.0212 - 1 = 26.8 percent, while the annual effective yield would be .02 x 12 = 24 percent. For a discussion of the
calculation of effective yield, see Rosenberg, Microcredit Interest Rates, CGAP Occasional Paper No. 1, revised
November 2002, pp. 5-10.
21
One approach to doing this can be found in Holtmann and Mommartz, Technical Guide for Analyzing the
Efficiency of Credit-Granting Non-Governmental Organizations, Saarbrucken 1996, pp. 106 ff.

Page ii